CAPER-α alternative splicing regulates the expression of vascular endothelial growth factor165 in Ewing sarcoma cells

Authors

  • Gangxiong Huang MD,

    1. Division of Pediatrics, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Zhichao Zhou MD,

    1. Division of Pediatrics, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Hua Wang PhD,

    1. Department of Cancer Biology, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Eugenie S. Kleinerman MD

    Corresponding author
    1. Division of Pediatrics, The University of Texas MD Anderson Cancer Center, Houston, Texas
    2. Department of Cancer Biology, The University of Texas MD Anderson Cancer Center, Houston, Texas
    • Division of Pediatrics, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit #87, Houston, TX 77030
    Search for more papers by this author
    • Fax: (713) 794-5042


Abstract

BACKGROUND:

TC-71 Ewing sarcoma cells overexpress vascular endothelial growth factor (VEGF) with a shift from the 189 to the 165 isoform.

METHODS:

The effect of CAPER-α on the expression of the VEGF isoforms, tumor growth, and vessel density was analyzed after transfection of TC-71 cells with CAPER-α cDNA or siRNA.

RESULTS:

CAPER-α correlated inversely with the VEGF165/VEGF189 mRNA ratio. Up-regulation of CAPER-α resulted in decreased tumor growth, tumor vessel density, and chemotactic activity of the cell's supernatant. CAPER-α expression was regulated by EWS/FLI-1 through a protein-protein interaction.

CONCLUSIONS:

Increased VEGF165 expression is secondary to the down-regulation of CAPER-α by EWS/FLI-1. CAPER-α mediates alternative splicing and controls the shift from VEGF189 to VEGF165. Cancer 2012;. © 2011 American Cancer Society.

INTRODUCTION

Vascular endothelial growth factor (VEGF) A plays a pivotal role in tumor angiogenesis. VEGF isoforms are produced by alternative splicing. The 121, 165, and 189 isoforms predominate in cancer cells1, 2 and differ in solubility, receptor affinity, and mitogenic potency. VEGF189 is membrane bound, with the highest affinity to heparin and heparan sulfates. It has less angiogenic activity in vivo than VEGF121 and VEGF165.3, 4 VEGF121 is freely diffusible with no binding to heparin. VEGF165 is also secreted but, unlike VEGF121, has heparin-binding ability and binds to neuropilin, resulting in increased mitogenic effects.5 VEGF165 expression correlates with increased tumor vascularity and a poorer prognosis in osteosarcoma patients.6

We demonstrated that Ewing sarcoma cells overexpress VEGF with a shift in isoform expression from the 189 to the 165 isoform.7, 8 VEGF165 was critical for Ewing sarcoma tumor vessel development and growth in vivo and could not be replaced by VEGF189.8, 9 Molecular cloning of the cDNA for VEGF revealed that VEGF isoform expression was controlled by alternative splicing of different exons,1 which in turn selectively joined different protein coding elements.10 The regulation and expression of the 3 isoforms and what causes a shift from the 189 to 165 isoform have not been defined.

CAPER-α, a transcriptional coactivator of AP-1, ERα, and ERβ,11 alters the ratio of VEGF121 and VEGF189 in breast cancer cells.12 We now demonstrate that CAPER-α mediates alternative splicing and controls the shift from VEGF189 to VEGF165 in Ewing cells. Down-regulation of CAPER-α resulted in an increase in the ratio of VEGF165 to VEGF189 (VEGF165/VEGF189). Transfection of CAPER-α into TC-71 cells lowered the VEGF165/ VEGF189 ratio and led to decreased tumor vessel density and growth in vivo. We also demonstrate that EWS/FLI-1 inhibits CAPER-α expression mediated by protein-protein interaction.

MATERIALS AND METHODS

Plasmids

pSilencer 2.1-U6 hygro plasmid (Ambion, Austin, Tex) was used to construct vector-expressing hairpin small interfering RNA (siRNA) to suppress CAPER-α expression.9, 12CAPER-α cDNAs were isolated and amplified from TC-71 cells by real time polymerase chain reaction (RT-PCR).11 PCR primers were 5′-CTCGGATCCGGAATGGCAGACG-3′ (forward) and 5′-CCTCTAGATCATCGTCTACTTGGAAC-3′ (reverse), with BamHI and XbaI recognition sites on the ends. The CAPER-α–GFP vector, with GFP attached to the COOH terminus, was constructed by subcloning the full-length CAPER-α cDNA fragments into the pEGFP-C1 plasmid (Clontech, Mountain View, Calif). The plasmid expressing c-Myc peptide epitope-tagged CAPER-α was constructed by subcloning of the full-length CAPER-α cDNA fragments into the pcDNA 3.1 vector (Invitrogen, Carlsbad, Calif).

CAPER-α–pLV was constructed by inserting cytomegalovirus promoter-driven CAPER-α cDNA into the 3′ long terminal repeat of pWPXL, a lentivirus vector that contains the internal ribosome entry site driving GFP expression, as described previously.13 pLV vector was the control. EWS/FLI-1 siRNAs were described previously.14 All the constructs were confirmed by DNA sequencing. Plasmids expressing M2-Flag–tagged EWS/FLI-1, EWS, or FLI-1 were provided by Dr. Liu Yang (the University of Washington School of Medicine, Seattle, Wash).

Cell lines, Transfection, and Transduction

Human embryonic kidney cells (293T; American Type Culture Collection [ATCC], Manassas Va); human Ewing sarcoma cells TC-71,14 A4573,14 RD-ES (ATCC), and SK-ES (ATCC); mouse lung vessel endothelial cells (Dr. Robert Langley, The University of Texas MD Anderson Cancer Center, Houston, Tex); and murine bone marrow (BM)-derived mesenchymal stem cells (MSCs) were cultured as previously.14, 15 Normal human osteoblasts (Clonetics, San Diego, Calif) were cultured in osteoblast growth medium (Clonetics) containing 10% FBS and 100 μg/mL ascorbic acid. All cells were tested for Mycoplasma by RT-PCR and verified to be free of pathogenic murine viruses (National Cancer Institute–Frederick Cancer Research & Development Center, Frederick, Md).

Transient transfection was performed for 48 hours with LipofectAMINE 2000 (Invitrogen). Stable single-cell colonies of TC-71 cells were selected after incubation in 400 μg/mL hygromycin B. Recombinant lentiviruses, CAPER-α–pLV or pLV, were produced by transient transfection of 293T cells13 and harvested 24 hours after transfection. For transduction, TC-71 cells were plated on 6-well plates (5 × 105cells/well) and then incubated overnight in medium containing recombinant lentivirus vectors. GFP-positive cells were sorted and cultured for in vivo experiments.

RNA and Protein Analysis

Total RNA was isolated and purified from cultured cells or homogenized tumor tissues using Trizol Reagent (Invitrogen). Reverse transcription was performed with oligo-dT primer (Promega, Madison Wis), followed by PCR using an iTaq DNA polymerase kit (Bio-Rad Laboratories, Hercules, Calif) with specific PCR primers for VEGF isoforms,16 VEGF165b,17 EWS/FLI-1(forward, 5′-GCCTCCTATGCAGCTCAGTC-3′; reverse, 5′-GGT TGTAACCCCCTGTGCTA-3′), and VEGF receptor-2 (VEGFR-2; forward, 5′-GTGACCAACATGGAGTCG TG-3′; reverse, 5′-TGCTTCACAGAAGACCATGC-3′). The glyceraldehyde-3-phosphate dehydrogenase mRNA level was the internal control. Quantification of PCR results was performed using Scion imaging software (Scion Corporation, Washington, DC).

VEGF mRNA isoforms were quantified by RT-PCR using an iQ SYBR Green Supermix (Bio-Rad Laboratories) and previously described PCR primers.18 Each DNA sample was analyzed by 3 replicate assays. β-Actin mRNA was quantified for normalization.

Nuclear protein was extracted from cultured cells at 80% confluence or 48 hours after transfection. Cells were lysed in Buffer A (10 mM HEPES [pH 7.9], 2.5 mM MgCl2, 100 mM NaCl, 0.5 mM dithiothreitol, and 1× protease inhibitor [Calbiochem, San Diego, Calif]) at 4°C until >90% cells were trypan blue (Sigma-Aldrich, St Louis, Mo) positive. After centrifugation (12,000 × g at 4°C), the nuclear pellets were resuspended in Buffer C (20 mM HEPES [pH 7.9], 25% glycerol, 1.5 mM MgCl2, 450 mM NaCl, 0.2 mM ethylenediaminetetraacetic acid, 0.5 mM dithiothreitol, and 1× protease inhibitor) and then incubated for 30 minutes at 4°C, followed by centrifugation. Supernatants were collected as nuclear proteins. Tumor tissue lysates were homogenized in lysis buffer (20 mmol/L HEPES [pH 7.4], 1% Triton X-100, 10% glycerol, and 1× protease inhibitor) and then centrifuged (12,000 × g at 4°C). The protein was denatured, loaded onto a 10% sodium dodecyl sulfate-polyacrylamide gel, and then transferred to a nitrocellulose membrane (Amersham, Piscataway, NJ). Western blot analysis using antihuman HCC1 (CAPER-α) antibody (Abcam, Cambridge, Mass or Bethyl Laboratories, Montgomery, Tex), anti-TATA box binding protein antibody (Abcam), antihuman FLI-1 polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, Calif), and β-actin antibody (Sigma Chemical, St Louis, Mo) was performed using enhanced chemiluminescence (Amersham). EWS/FLI-1 protein (∼68 kD) was detected by the anti–FLI-1 antibody and was distinguished from FLI-1 protein (∼54 kD) by its molecular weight.19 Quantification of results was performed using Scion imaging software.

For coimmunoprecipitation, after precleaning with mouse immunoglobulin G (IgG), 0.2 mL of nuclear proteins were incubated with anti–c-Myc monoclonal antibody (Theroma Fisher Scientific, Fremont, Calif) overnight. A/G plus agarose (Santa Cruz Biotechnology) was then added and incubated for an additional 3 hours at 4°C. After 3 washes with NP-40 buffer (50 mM Tris [pH 7.5], 150 mM NaCl, and 0.05% NP-40), the immunoprecipitates were blotted with mouse anti–Flag-M2 monoclonal antibody (Sigma-Aldrich) for Western blot analysis.

Migration Assay

Mouse lung vessel endothelial cells (1 × 105) in transwells were inserted into 24-well plates containing 600 μL of cultured supernatants and incubated at 37°C for 8 hours. Migrated cells were fixed, stained with hematoxylin & eosin, and quantified in 5 random high-power fields.

In Vivo Model

Four- to 5-week-old specific pathogen-free athymic nude mice (Charles River Breeding Laboratories, Wilmington, Mass) were maintained in an animal facility approved by the American Association for Laboratory Accreditation. Animal care protocols were approved by the institutional animal care and use committee. TC-71 cells transduced with lentiviruses CAPER-α–pLV (TC-CAPER-α–pLV cells) or pLV (TC-pLV cells) were injected subcutaneously.7 Expression of CD31 and VEGF isoforms in tumor tissues was analyzed by immunohistochemistry and RT-PCR.

Immunohistochemistry and Microscopy

TC-71 cells were fixed in cold acetone and blocked with 4% fish gelatin in phosphate-buffered saline (PBS). Immunohistochemical staining was performed using primary antihuman HCC1 (CAPER-α) and anti-EWS antibodies. Cells were incubated with primary antibodies overnight, washed in PBS, and incubated with 4% fish gel before the addition of secondary fluorescent antibodies. Cy5-labeled antimouse secondary antibodies (Jackson Immunoresearch, West Grove, Pa) and Alexa488-labeled antirabbit secondary antibodies (Molecular Probes, Carlsbad, Calif) were used. Nuclei were labeled with Sytox Green (Jackson Immunoresearch). Confocal microscopy images were captured using a Zeiss Laser Confocal Microscope (Carl Zeiss MicroImaging, Thornwood, NY) and LSM software (Zeiss).

Statistical Analysis

Analysis was done using analysis of variance and the 2-tailed Student t test; P < .05 was considered statistically significant.

RESULTS

CAPER-α Expression Is Inversely Correlated with VEGF165

TC-71 cells express higher levels of VEGF165 and minimal or no VEGF189 compared with normal human osteoblast cells.7 To determine whether CAPER-α expression contributes to this isoform shift, we analyzed the expression of CAPER-α in TC-71 cells. Normal human osteoblast and MSCs were the controls. CAPER-α expression was significantly decreased in TC-71 cells (Fig. 1A) and was inversely correlated with the level of VEGF165 (Fig. 1B). TC-71 cells showed lower levels of VEGF189 and a higher VEGF165/VEGF189 ratio than normal human osteoblasts cells as calculated by RT-PCR (Fig. 1C). Three other Ewing sarcoma cell lines were also analyzed. The level of CAPER-α expression was lowest in TC-71 cells and highest in RD-ES cells (Fig. 1D). As anticipated, RD-ES cells showed higher levels of VEGF189 and lower levels of VEGF165, yielding a lower VEGF165/VEGF189 ratio than that in TC-71 cells (Fig. 1E, F). SK-ES and A4573 cells also had higher levels of CAPER-α and lower VEGF165/VEGF189 ratios than TC-71, confirming an inverse correlation between CAPER-α and VEGF165. Thus, CAPER-α expression may play a critical role in controlling the expression of VEGF165 and the ratio of VEGF165 to VEGF189.

Figure 1.

Expression of CAPER-α and the vascular endothelial growth factor (VEGF) isoforms is shown in various cell lines. Nuclear protein and total RNA were extracted from various cell lines. CAPER-α was detected by Western blot analysis (A and D). TATA-binding protein (TBP) was the internal nuclear protein control. The CAPER-α ratio was calculated by comparing the expression level (density) of CAPER-α to that of TBP. Expression of VEGF isoforms was determined by reverse transcription polymerase chain reaction (B and E). The VEGF165/VEGF189 ratio was calculated by comparing the expression level (density) of VEGF165 to that of VEGF189. Real time polymerase chain reaction was also used to calculate the VEGF165 of 189 ratio (C and F). Bars represent the standard deviation. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; MSC, mesenchymal stem cells; NHOST, normal human osteoblasts.

CAPER-α Alters VEGF165/VEGF189 mRNA Ratio

To determine whether the inhibition of CAPER-α affected the expression of VEGF isoforms, a pSilencer plasmid producing CAPER-α siRNA was transfected into TC-71 cells. As measured by Western blot analysis (Fig. 2A), CAPER-α expression in 2 of the stable clones was decreased (TC-siCAPER-α 5 and 11), but not in the TC-siCAPER-α 2 clone. The decreased CAPER-α expression in TC-siCAPER-α 5 cells correlated with an increase in VEGF165 mRNA (Fig. 2B) and an increase in the VEGF165/VEGF189 ratio (Fig. 2C), compared with those in parental TC-71 and TC-si control cells. Similarly, CAPER-α expression in A4573 cells was decreased after transient transfection with CAPER-α siRNA, resulting in an increase in the VEGF165/VEGF189 ratio (Fig. 2D). Cultured supernatants harvested from TC-siCAPER-α 5 cells induced an increase in the migration of mouse-derived endothelial cells (mouse lung vessel endothelial cells) that express VEGFR-2, compared with supernatants harvested from parental and TC-si control cells (Fig. 2E). Similar results were seen with the TC-siCAPER-α 11 cells and A4573 cells transfected with Caper-α siRNA (data not shown).

Figure 2.

Suppression of CAPER-α increases vascular endothelial growth factor (VEGF)165 and the VEGF165 of 189 ratio. (A) TC-71 cells were transfected with CAPER-α siRNA. Suppression of CAPER-α in the stable colonies (TC-siCAPER-α 2, 5, and 11) was confirmed by Western blot analyses. (B) VEGF isoforms were quantified by reverse transcription polymerase chain reaction. The VEGF165/VEGF189 ratio was calculated by comparing the expression level of VEGF165 to that of VEGF189. (C) VEGF165/VEGF189 ratio was also calculated using real time polymerase chain reaction. (D) A4573 cells were transfected with CAPER-α siRNA for 48 hours. CAPER-α expression was confirmed by Western blot analyses. Real time polymerase chain reaction was used to determine the VEGF165/VEGF189 ratio. (E) The cultured supernatants were collected from the different cells and assayed for chemotactic activity using a migration assay for mouse lung vessel endothelial cells (MLVEC). Results are expressed as the average number of migrated cells in 5 random high power fields ± the standard deviation. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; TBP, TATA-binding protein.

Transfection of TC-71 cells with CAPER-α–GFP (TC-CAPER-α–GFP) resulted in expression of GFP and overexpression of CAPER-α. Control GFP-transfected cells expressed GFP but no CAPER-α (Fig. 3A). GFP expression in the TC-CAPER-α–GFP cells was localized to the nuclei, whereas transfected GFP control cells showed GFP expression in both nucleus and cytoplasm (Fig. 3B). Up-regulation of CAPER-α resulted in a decrease in VEGF165 expression, an increase in VEGF189 expression (Fig. 3C), a decrease in the VEGF165/VEGF189 mRNA ratio (Fig. 3C, D), and a decrease in the chemotactic activity of the cultured supernatant (Fig. 3E), compared with transfected GFP control cells. As detected by enzyme-linked immunosorbent assay (ELISA), there was a 25% reduction in secreted VEGF protein levels in the supernatant from TC-CAPER-α–GFP cells as compared with that in control transfected GFP cells (P < .001). This is consistent with the decrease in VEGF165 expression, as VEGF189 is primarily membrane bound. These results indicate that up-regulation of CAPER-α decreased both VEGF165 expression and protein production as well as the chemotactic activity of the cultured supernatant for endothelial cells.

Figure 3.

Up-regulation of CAPER-α decreases vascular endothelial growth factor (VEGF)165 and the VEGF165/VEGF189 ratio. TC-71 cells were transiently transfected with pEGFP–CAPER-α (TC-CAPER-α–GFP) or control pEGFP (TC-GFP) cDNA. (A) Expression of CAPER-α was confirmed by Western blot. (B) Expression of GFP was assessed by microscopy. In TC-CAPER-α–GFP cells, GFP expression was localized in the nuclei (a: original magnification, ×40; c: original magnification, ×200). GFP was expressed in both cytoplasm and nuclei in the control transfected TC-GFP cells (b: original magnification ×40; d: original magnification, ×200). (C) Expression of VEGF isoforms was analyzed by reverse transcription polymerase chain reaction. (D) Real time polymerase chain reaction was used to determine the VEGF165/VEGF189 ratio. (E) Supernatants were collected separately from the different cultured cells and analyzed for chemotactic activity using mouse lung vessel endothelial cells (MLVEC). Results are expressed as the average number of migrated cells in 5 random high power fields ± the standard deviation.

Up-Regulation of CAPER-α Expression Inhibits TC-71 Tumor Formation In Vivo

VEGF165 is critical for TC-71 growth in vivo.7-9 Selective inhibition of VEGF165 resulted in the suppression of tumor growth, which could not be rescued by VEGF189.8 We investigated the effect of up-regulation of CAPER-α expression on TC-71 tumor growth in vivo. TC-71 cells were transfected with CAPER-α–pLV or pLV control vector. Increased CAPER-α expression was seen in cells after transfection with CAPER-α-pLV (Fig. 4A). TC-CAPER-α–pLV or TC-pLV cells were injected subcutaneously. Tumor formation (Fig. 4B) and tumor vessel density (Fig. 4C, D) after the subcutaneous injection of TC-CAPER-α–pLV cells were significantly inhibited compared with the TC-pLV control cells. RNA was extracted from the excised tumors at the time of sacrifice and then analyzed by RT-PCR for VEGF165 expression. As expected, TC-CAPER-α–pLV tumors showed decreased VEGF165 compared with control TC-pLV tumors (Fig. 4E).

Figure 4.

Effects of CAPER-α on tumor growth and tumor vessel density in vivo are shown. TC-71 cells were transduced with CAPER-α–pLV or pLV alone. (A) Increased CAPER-α expression was confirmed by Western blot analyses. (B) Tumor growth was monitored after subcutaneous injection of TC-CAPER-α–pLV or TC-pLV cells. (C) Representative CD31 staining (red), Hoechst staining (blue), and GFP expression are shown in tumor tissue sections. (D) Mean tumor vessel density was calculated by counting the number of CD31-positive blood vessels in 5 random microscopic fields ± standard deviation. (E) Expression of vascular endothelial growth factor (VEGF)165 in the TC-CAPER-α-pLV or TC-pLV tumors was assayed by reverse transcription polymerase chain reaction. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as the loading control. (F) VEGFR-2 and VEGF165b were assayed by reverse transcription polymerase chain reaction after transduction of TC-71 cells with pLV control or CAPER-α-pLV. TBP, TATA-binding protein.

The suppression in tumor formation could not be explained by a change in the doubling time of the TC-CAPER-α cells (29.6 hours) versus the TC-pLV cells (26.1 hour, P = .36), a decrease in VEGFR-2 (Fig. 4F), or the increased expression of VEGF165b (Fig. 4F), an antiangiogenic isoform of VEGF165.17 These data indicate that the effect of CAPER-α on tumor growth was secondary to the alteration of VEGF165.

Effect of EWS/FLI-1 on CAPER-α Expression

EWS/FLI-1 regulates several downstream genes, including VEGF.20, 21 Because of the link between EWS/FLI-1 and VEGF20, 21 and our finding that the expression of CAPER-α inversely correlates with the expression of VEGF165, we investigated whether EWS/FLI-1 regulated CAPER-α expression. TC-71 cells were transfected with EWS/FLI-1 siRNA. TC-siEWS/FLI-1 cells showed decreased expression of EWS/FLI-1 and increased CAPER-α expression (Fig. 5A). Similarly, A4573 cells were transiently transfected with EWS/FLI-1 siRNA to decrease expression of EWS/FLI-1, resulting in increased expression of CAPER-α (Fig. 5B). By contrast, transfection of EWS/FLI-1 into MSCs using cDNA transfection led to decreased CAPER-α expression (Fig. 5B). We found evidence of a protein-protein interaction between CAPER-α and EWS/FLI-1 in TC-71 cells as determined by immunoprecipitation (Fig. 5C). These results suggest that EWS/FLI-1 down-regulated CAPER-α expression. To confirm this interaction, a coimmunoprecipitation assay was performed. HEK293 cells were cotransfected with a plasmid expressing c-Myc–tagged CAPER-α and another plasmid expressing Flag-tagged EWS/FLI-1, EWS alone, FLI-1 alone, or a neo control. HEK293 cells were also transfected with a plasmid expressing Flag-tagged EWS/FLI-1, EWS, FLI-1 alone, or a neo control. After immunoprecipitation with anti–c-Myc antibody, immunoblotting with anti-Flag antibody demonstrated that only EWS/FLI-1 and EWS proteins were detected in HEK293 cells cotransfected with c-Myc–tagged CAPER-α and Flag-tagged EWS/FLI-1 or EWS, but not in the cells only transfected with Flag-tagged EWS/FLI-1 or EWS, indicating that CAPER-α interacted with EWS/FLI-1 and EWS, but not with FLI-1 (Fig. 5D) in HEK293 cells. This interaction was further confirmed by the colocalization of the EWS and CAPER-α in TC-71 cells (Fig. 5E). Furthermore, transfection of EWS/FLI-1 into MSCs led to increased VEGF165 expression as assessed by RT-PCR (Fig. 5F). These results suggest that EWS/FLI-1 decreased CAPER-α expression by a mechanism that involves a protein-protein interaction in TC-71 cells. This down-regulation of CAPER-α may in turn result in increased VEGF165 expression and an increased VEGF165/VEGF189 ratio.

Figure 5.

EWS/FLI-1 interacts with CAPER-α. (A) TC-71 cells were transiently transfected twice, 48 hours apart, with EWS/FLI-1 siRNA. CAPER-α and EWS/FLI-1 expression were quantified by Western blot analysis. (B) A4573 cells were transiently transfected for 48 hours with EWS/FLI-1 siRNA. EWS/FLI-1 and CAPER-α expression were quantified by Western blot analysis. Mesenchymal stem cells (MSCs) were transfected with EWS/FLI-1 (MSC–EWS/FLI-1) or vector only (MSC-neo). CAPER-α and EWS/FLI-1 expression were determined by Western blot analysis. (C) Cell lysates prepared from TC-71 were immunoprecipitated with either anti–CAPER-α antibody or immunoglobulin G (IgG) control. The immunoprecipitates as well as cell lysates without immunoprecipitation as inputs were then blotted with the anti–FLI-1 antibody. Cell lysates of LM-7 osteosarcoma cells were used as the negative control, because these cells do not express EWS/FLI-1. (D) HEK-293T cells were cotransfected with plasmids expressing the c-Myc peptide epitope-tagged CAPER-α and plasmids expressing Flag-tagged EWS/FLI-1, EWS, FLI-1, or neo vector control, or were only transfected with plasmids expressing Flag-tagged EWS/FLI-1, EWS, FLI-1, or neo vector control. The cell lysates were first incubated with anti–c-Myc antibody to immunoprecipitate c-Myc–tagged CAPER-α or incubated with nonimmune IgG as a control. The immunoprecipitates were then blotted with anti–Flag M2 monoclonal antibody to detect Flag-tagged proteins. Total cell lysates, without immunoprecipitation, were also blotted with anti–Flag M2 monoclonal antibody as an input control. (E) TC-71 cells were analyzed by immunohistochemistry using anti-CAPER-α (red) and anti-EWS (green) antibodies. Yellow staining depicts colocalization between EWS and CAPER-α. Nuclei staining is shown in blue. (F) MSCs were transfected with EWS/FLI-1 (MSC–EWS/FLI-1) or vector only (MSC-neo). VEGF and EWS/FLI-1 expression was determined by reverse transcription polymerase chain reaction. The ratio of vascular endothelial growth factor (VEGF)165 to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was quantified. IB, immunoblots; IP, immunoprecipitates; TBP, TATA-binding protein.

DISCUSSION

Expression of CAPER-α, a transcriptional coactivator for steroid receptors, was inversely correlated to the expression of VEGF165 and the VEGF165/VEGF189 ratio in 3 different Ewing sarcoma cell lines. Increasing or decreasing CAPER-α expression resulted in a change in VEGF165 expression and in the chemotactic activity of cultured tumor cell supernatants for endothelial cells. The antibody used in the commercial ELISA kit to quantify VEGF165 may also bind to the 121 isoform. Although we did not detect a change in VEGF121 expression, some of the decreased chemotactic activity of the cultured supernatants from the TC-CAPER-α-GFP cells may be because of a reduction in VEGF121 in addition to VEGF165. Changing CAPER-α expression affected the phenotypic behavior of TC-71 cells in vivo. CAPER-α–transfected TC-71 cells had decreased VEGF165, tumor formation, and tumor vessel density. This decrease in endothelial migration, tumor formation, and tumor vessel density was not secondary to a decrease in cell doubling time, reduced expression of VEGFR-2, or the induction of elevated levels of VEGF165b, an antiangiogenic isoform of VEGF165.17 The cell doubling time of control and CAPER-α-transfected cells was not statistically different. Transfection of CAPER-α had no effect on either VEGFR-2 or VEGF165b expression.

These findings identify a mechanism by which Ewing sarcoma cells can alter the microenvironment to optimize the formation of the new vasculature required for tumor growth and metastasis. Solid tumors must have an extensive vasculature network to bring in needed oxygen and nutrients. Without vasculature expansion, tumors remain small. Tumor cells that can stimulate the microenvironment to form vessels clearly have an advantage. Therefore, understanding the molecular pathways that control this process can lead to identifying and designing specific targeted therapy to interfere with this tumor-microenvironment interaction.

VEGF is a key protein involved in tumor vessel expansion. Elevated VEGF has been shown in several different tumors, including Ewing sarcoma, and is associated with increased tumor vessel density and poor patient outcomes.22-25 The VEGF gene is alternatively spliced, yielding different isoforms, including VEGF165 and VEGF189.1 VEGF189 is tightly bound to proteoglycans in the extracellular matrix or on the cell surface, whereas VEGF165 is soluble and secreted. The 165 isoform is a major chemoattractant for BM progenitor endothelial cells. By contrast, the 189 isoform has little or no chemotactic activity for BM cells. We have previously demonstrated that BM stem cells, which include endothelial precursors, play an important role in the formation of new Ewing sarcoma vessels, and that VEGF165 is responsible for both the directed migration and differentiation of these BM stem cells into endothelial cells that make up the tumor vessels.7, 8, 26-28 The selective inhibition of VEGF165 resulted in decreased BM cell migration into the tumor and small tumors, with significantly fewer vessels.8, 9 VEGF189 could not compensate for the loss of VEGF165.8 These data indicate that there is a selective advantage for tumor cells to produce VEGF165 as opposed to VEGF189. Varying the expression of these different isoforms can therefore affect the speed and efficiency of vascular expansion.

VEGF mRNA is transcribed from 8 exons and then alternately spliced to generate the different isoforms. The 165 and 189 isoforms differ in their incorporation of exons 6 and 7 of the full-length gene. These exons encode a cationic domain that confers heparin-binding activity. VEGF189 contains both exons 6 and 7, whereas VEGF165 contains only exon 7. Inclusion of both exons 6 and 7 gives VEGF189 an increased charge, resulting in greater extracellular matrix association (relative to VEGF165) and its membrane-bound form. Because it is membrane bound and not secreted, VEGF189 has little impact on stimulating the migration of BM and endothelial precursors to the neovascular area, a function that is critical to the formation of the Ewing sarcoma vascular network.7, 8, 27, 28 Thus, there is an advantage to tumor cells producing VEGF165.

These data suggest that CAPER-α regulates the alternative splicing of VEGF in favor of the 189 isoform and that its down-regulation results in increased VEGF165. Manipulation of this gene alone affected tumor vascular formation as assessed by tumor vessel density and tumor growth. Given that VEGF189 (as opposed to VEGF165) includes the protein fragment coded by exon 6, we hypothesize that CAPER-α may target exon 6 for mRNA expression during the alternative splicing process (Fig. 6). Increasing exon 6 inclusion in turn favors VEGF189 expression and alters or lowers the VEGF165/VEGF189 ratio.

Figure 6.

Proposed regulation of vascular endothelial growth factor (VEGF) alternative splicing by CAPER-α is shown. The precursor mRNA of VEGF includes 8 exons (box, E1–E8) interspersed by noncoding introns (solid line). CAPER-α induces alternative splicing and inclusion of exon 6 (dotted line indicates joining of exons) by the spliceosome, resulting in the expression of VEGF189. EWS/FLI-1 prevents this action, resulting in the selective exclusion of exon 6.

In addition to being membrane bound, there is a difference in the biological function of the 189 and 165 isoforms.29-31 Overexpression of VEGF165 in human melanoma cells, which were initially VEGF negative, resulted in aggressive tumor growth in vivo, whereas parental cells and those transfected with VEGF189 were nontumorigenic and dormant when injected into nude mice.30 These VEGF165-transfected cells had a dense vascular network with effective tumor perfusion. Another study showed that only VEGF165 (not VEGF189 or VEGF121) was able to rescue VEGF-inhibited tumors. Tumors expressing VEGF189 showed small, convoluted, and inadequately perfused vessels compared with those in tumors expressing only VEGF165. These studies and our previous results showing the importance of BM cell migration to the formation of the Ewing tumor vasculature support our contention that there is a selective advantage in terms of vascular development afforded to tumors that express greater amounts of the 165 isoform.

Our data also indicate a possible mechanism for how CAPER-α is regulated in Ewing cells. Ewing sarcoma tumors are characterized by a unique chromosomal translocation between chromosomes 11 and 22 that leads to the formation of fusion genes encoding proteins composed of the transcriptional domain of EWS and the binding domain of 1 of 5 ETS transcription factors. The most common translocation is between the EWS and FLI genes resulting in production of the EWS/FLI-1 fusion protein, which functions as an aberrant transcription factor.32 Several EWS/FLI-1 target genes have been identified and implicated in the initiation and progression of Ewing sarcoma.33-35 The EWS/FLI-1 protein has been shown to be a transcription factor for VEGF.21 We show here that, in addition to regulating expression and production of VEGF, EWS/FLI-1 may also control the alternative splicing process for VEGF in favor of VEGF165 by regulating the expression of CAPER-α.

EWS/FLI-1 interacts with SF1, U1 small nuclear ribonucleoprotein (U1C), and hyperphosphorylated RNA polymerase II, contributing to cellular transformation by affecting RNA splicing.19, 36, 37 Here we have demonstrated that inhibiting EWS/FLI-1 in TC-71 or A4573 cells led to increased CAPER-α. By contrast, transfecting EWS/FLI-1 into MSC cells resulted in decreased expression of CAPER-α. Furthermore, we demonstrated that EWS/FLI-1 has a protein-protein interaction with CAPER-α. These data support the hypothesis that EWS/FLI-1 is involved in the alternative splicing process of VEGF through the regulation of CAPER-α.

The difference in CAPER-α expression among the 3 different Ewing cell lines may be explained by the finding of 3 different subtypes of EWS/FLI-1 fusion proteins, which result from different fusions between the N terminal of EWS and the C terminal of FLI.38 TC-71 cells express the type 1 fusion, whereas RD-ES cells express the type 2 fusion.38 Differences in the fusion proteins may in turn affect their ability to regulate CAPER-α. We anticipate that the type 1 fusion protein found in TC-71 cells would be more effective at down-regulating CAPER-α than the type 2 fusion protein.

VEGF165 plays a central role in the growth of Ewing sarcoma and the expansion of its vascular network. The overexpression of VEGF165 in some Ewing tumors may be mediated by a switch from the 189 isoform to the 165 isoform, which is secondary to the down-regulation of CAPER-α. Exon 6 is part of the VEGF189 mRNA but is not included in the VEGF165 mRNA. Therefore, the shift from the 189 to the 165 isoform may be mediated by the exclusion of exon 6 from the VEGF precursor mRNA because of the down-regulation of CAPER-α. EWS/FLI-1 may control the transcription of CAPER-α, resulting in its down-regulation in Ewing sarcoma cells.

FUNDING SOURCES

Supported by National Cancer Institute CA103986 (E.S.K.) and CA 16672 Cancer Center Support Core Grant.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary